The 60-second answer
Hosted call control makes location irrelevant. Extensions in two offices and six homes all register to the same platform, dial each other by short number, and share one auto-attendant. No site-to-site link, no VPN, one extension band. Pricing is by extension count, not by building.
The assumption worth dropping
For most of the history of business telephony, a phone system was a box in a closet, and the box was in a building. Two buildings meant two boxes. Connecting them meant a dedicated circuit between the sites, which was expensive enough that most businesses simply did not, and lived with two separate phone systems that happened to belong to the same company.
That constraint is gone, and a surprising number of organisations still budget as though it were not. When call control is hosted, an extension registers to a platform over the internet [1]. Whether the device sits in the Moncton office, the Saint John office, or someone's kitchen in Sackville is a networking detail with no architectural consequence. Extension 214 is extension 214.
Two offices do not need two phone systems. They need one phone system and two internet connections.
What one dial plan actually gives you
- Short dialling between sites. Anyone dials 214 and reaches that person, regardless of which building either party is in. No area code, no transfer through reception, no long distance.
- Transfers that work across locations. A call answered in one office transfers to an extension in another exactly like a transfer down the hall — blind or attended, with the same keypresses.
- One auto-attendant, or several. A single main number can offer a location choice, or each site can keep its own number and its own greeting, both landing on the same platform.
- Departments that span sites. A support group can ring three phones in one city and two in another simultaneously. Callers cannot tell, and should not need to.
- One directory. Names and extensions in one place, dialled by name from any handset.
- One set of changes. A holiday closure is configured once, not once per building.
The networking question people ask first
"Do I need a connection between my offices?" No. This is the single most common misconception carried over from the tie-line era.
Each site needs adequate internet for its own phones — roughly 100 kbps per concurrent call in each direction, plus voice prioritisation on the local router. That is it. Both sites reach the hosted platform independently; they never need to reach each other. Two offices in different provinces with different providers are no harder to run on one dial plan than two floors of the same building.
The practical consequence is that adding a location is not a project. A new office with eight phones plugs those phones in, they register, and they are on the dial plan. There is no circuit to order and no lead time beyond the internet connection itself.
Remote workers, without a VPN
Remote extensions are where hosted call control separates most sharply from an on-premise system. With a PBX in your closet, giving someone a home extension means exposing that server to the internet, which is exactly the kind of internet-facing service that turns into toll fraud when it is not maintained.
With hosted call control there is nothing in your office to expose. A remote phone registers directly to the platform, and browser-based softphones use standard web transport for signalling [2], meaning a laptop with a headset works from any network without special client software or a tunnel back to head office.
That does not make remote endpoints risk-free, and federal telework guidance is worth reading rather than assuming [3]. The realistic controls are unglamorous:
- Strong, unique credentials per extension — never shared or reused across staff.
- Extensions removed the day someone leaves, not at the next audit [4].
- Awareness that a call taken at home is subject to the same safeguards obligation as one taken at a desk, which is a conversation about who else is in the room, not a technical control [5].
- Voicemail delivered to work email rather than personal accounts.
Designing the numbering plan before you build it
The one decision worth making deliberately is how extension numbers are allocated, because changing it later means reprinting cards and retraining everyone.
Two approaches work. The first is site-prefixed: extensions at the first location run 2xx, the second 3xx, the third 4xx. Anyone hearing an extension knows immediately which building it belongs to, which helps reception and helps callers who ask to be put through to a specific office. The cost is that moving a person between sites means changing their number.
The second is location-agnostic: numbers are assigned sequentially as people join, and nobody's extension encodes where they sit. This suits organisations where staff move between sites, work hybrid, or where the physical location genuinely does not matter to callers. It is the better default for most businesses now, precisely because it stops treating a desk as the unit of identity.
Whichever you choose, reserve blocks rather than filling numbers sequentially from 200. Keep 2xx for people, 6xx for shared and common-area phones, 8xx for analogue devices like the fax and the elevator, and 9xx for departments and hunt groups. When someone dials a number and hears something unexpected, that structure tells you what kind of thing they reached without looking anything up.
Also decide early what an unanswered extension does at each site. It is entirely reasonable for the second office to roll to a group at the first after four rings during the hours it is thinly staffed, and equally reasonable for it to go to voicemail. What is not reasonable is discovering the answer by accident six months in, which is what happens when this is left at the default.
What it costs
Pricing is by extension band, not by location. Two offices and a set of remote staff totalling 30 extensions sit in the $399 band exactly as a single 30-person office would.
| Two separate systems | One dial plan | |
|---|---|---|
| Platform cost | 2 × entry-band minimum | One band covering both |
| Spare handsets | Duplicated per site | Shared pool |
| Site-to-site calls | External, dialled in full | Four-digit, internal |
| Configuration changes | Twice | Once |
| Adding a third location | New system | Plug phones in |
Bands are $99, $219, $399 and $699 CAD per month at 10, 25, 50 and 100 extensions, plus HST. Setup is $350 at the entry band ($250 with a 12-month commitment), rising to $750, $1,200 and $1,900. Multi-location operation is common enough among Canadian businesses [6] that treating it as an exception no longer reflects how most companies work.
Where separate systems still make sense
Two cases, honestly. If the locations are genuinely different businesses — different brands, different owners, no shared staff — separate systems match the reality and keep billing clean. And if one site has poor or unreliable internet, an on-premise system there with its own analogue fallback may be the more robust answer, with the other sites hosted. Hybrid arrangements like that are normal and worth designing at a site visit rather than assumed away.
Bottom line
One phone system, however many buildings. Sites need internet, not links to each other. Remote staff need credentials, not VPNs [2][3]. And because pricing follows extension count rather than address, consolidating usually costs less than the arrangement it replaces — while making a third location a matter of plugging phones in.